Details

Delay and disruptive tolerant networks (sometimes also referred to as intermittently connected mobile networks) are networks where most of the time, there does not exist a complete end-to-end path from the source to the destination. Even if such a path exists, it may be highly unstable because of the topology changes due to mobility and may change or break soon after it has been discovered. This situation arises when the network is quite sparse, in which case it can be viewed as a set of disconnected, time varying cluster of nodes. Examples of such networks include vehicular ad hoc networks, sensor networks for wildlife tracking and habitat monitoring, military networks, deep-space inter-planetary networks, nomadic communities networks, networks of mobile robots, underwater networks etc.

Traditional mobile ad hoc routing protocols will fail for these networks because they require the existence of complete end-to-end paths to be able to deliver any data. To overcome this issue, mobility-assisted routing schemes have been proposed that often make a mobile node store and carry a message around, until an appropriate communication opportunity arises. We have studied a range of representative mobility-assisted routing strategies, including randomized strategies, utility-based strategies, schemes using multiple message copies, and hybrid approaches. We have proposed an analytical framework which allows us to study and compare all these schemes in a realistic network. Key features of our analytical framework include the use of realistic mobility models, which allow nodes to behave differently from other nodes and to preferentially move in some local areas, and the complete analysis of network contention caused by limited bandwidth and interference.

Based on our studies, using flooding-based ideas result in severe contention, whereas using a single copy per message yields large delays. In contrast, we find that using a small, fixed number of copies, and then routing each copy independently, yields surprisingly good performance under a wide range of scenarios. Clearly, to achieve a desirable performance, one needs to intelligently decide how many copies to distribute, how to choose good relay nodes for the copies, and how to efficiently route each copy in an independent fashion using the information available locally to each node.